UNITED STATES DISTRICT COURT EASTERN DISTRICT OF TEXAS MARSHALL DIVISION

UNITED STATES DISTRICT COURT EASTERN DISTRICT OF TEXAS MARSHALL DIVISION
DECLARATION OF DR. JIAN-WEI HUANG IN SUPPORT OF PLAINTIFF OPTIWAVE SEMICONDUCTOR CORP.'S OPENING CLAIM CONSTRUCTION BRIEF
I, Dr. Jian-Wei Huang, hereby declare as follows:
I. INTRODUCTION AND IDENTIFICATION
1.  My name is Dr. Jian-Wei Huang. I am over the age of twenty-one and am competent to make this Declaration. I have personal knowledge of the facts set forth herein, and if called as a witness, I could and would competently testify to these facts.
2.  I am the Chief Technology Officer of Optiwave Semiconductor Corp. ("OWS" or "Optiwave"). My business address is 4810 Bayside Technology Parkway, Suite 300, Santa Clara, California 95054. I have held the position of Chief Technology Officer at OWS since January 2014.
3.  I make this Declaration under penalty of perjury pursuant to 28 U.S.C. § 1746, in support of Plaintiff Optiwave Semiconductor Corp.'s Opening Claim Construction Brief ("Opening Brief") filed concurrently herewith. The purpose of this Declaration is to provide testimony regarding: (a) the prosecution history of the four United States patents asserted in this action; (b) the intended scope of certain disputed claim terms as understood by me, as a named inventor on each of the four asserted patents; and (c) the state of the art in the field of silicon photonics at the time each of the asserted patents was filed.
4.  I have reviewed OWS's Opening Claim Construction Brief, as well as the patent specifications, claims, and prosecution histories of each of the four asserted patents. The statements I make in this Declaration are based on my personal knowledge, my review of the relevant documents, and my expertise in the field of silicon photonics, as described more fully below.
II. QUALIFICATIONS AND BACKGROUND
A. Education.  
5.  I received a Bachelor of Science degree in Physics from Zhejiang University, Hangzhou, China, in 2002. My undergraduate thesis focused on the nonlinear optical properties of lithium niobate waveguides.
6.  I received a Master of Science degree in Electrical Engineering from the University of Michigan, Ann Arbor, in 2005. My master's research involved the design and fabrication of silicon-on-insulator (SOI) rib waveguides for integrated optical circuit applications.
7.  I received a Doctor of Philosophy degree in Electrical Engineering from the Massachusetts Institute of Technology in 2009, with a concentration in integrated photonics. My doctoral dissertation, entitled "Monolithic Integration of High-Performance Silicon Photonic Devices for Optical Interconnect Applications," was supervised by Professor Elena Rostova in the Photonic microsystems Group. My doctoral research centered on the development of novel silicon-based photonic integrated circuits for high-speed data communications, including ring-resonator modulators, germanium-on-silicon photodetectors, and waveguide-based wavelength multiplexers.
B. Professional Experience
8.  Following the completion of my doctoral degree, I joined Optiwave Semiconductor Corp. in September 2009 as a Senior Photonics Design Engineer, shortly after OWS was co-founded in 2009 by Dr. Raymond Patel and Dr. Sarah Chen. In that role, I was responsible for the design, simulation, and layout of silicon photonic integrated circuits for optical transceiver applications.
9.  In 2011, I was promoted to the position of Director of Photonics Engineering, where I led a team of engineers responsible for the development of OWS's first-generation silicon-photonics transceiver modules for data-center interconnect applications. I oversaw the development of OWS's foundational photonic integrated circuit platform, including waveguide structures, modulator designs, and grating coupler architectures.
10.  In January 2014, I was appointed Chief Technology Officer of OWS, a position I have held continuously since that date. As CTO, I am responsible for the overall technical direction of the company, including research and development strategy, product architecture, intellectual property development, and technology partnerships. I report directly to the Chief Executive Officer and serve as a member of OWS's executive leadership team.
11.  Over the course of my career, I have accumulated more than fifteen years of professional experience in the field of silicon-photonics integrated circuit design. My areas of technical expertise include, but are not limited to: silicon-on-insulator waveguide design and fabrication; high-speed electro-optic modulator architectures; germanium-on-silicon photodetector integration; coherent optical transceiver architectures for 100G, 400G, and 800G data rates; waveguide-based wavelength-division multiplexing and demultiplexing; grating coupler design for fiber-to-chip coupling; thermal management of photonic integrated circuits; and CMOS-compatible fabrication processes for silicon photonics.
12.  I am a named inventor on fifty-three issued United States patents and twelve pending United States patent applications in the field of silicon photonics and related technologies. I am also a named inventor on twenty-one corresponding foreign patents and applications.
13.  Importantly, I am a named inventor on all four of the United States patents asserted in this action:
(a) U.S. Patent No. 10,234,847 ("the '847 Patent"), entitled "Multi-Channel Silicon Photonic Waveguide Multiplexer with Low Insertion Loss";
(b) U.S. Patent No. 10,891,193 ("the '193 Patent"), entitled "High-Bandwidth Coherent Optical Transceiver Integrated on Silicon Substrate";
(c) U.S. Patent No. 9,678,562 ("the '562 Patent"), entitled "Method for Thermal Management of Photonic Integrated Circuits Using Embedded Micro-Channel Heat Sinks"; and
(d) U.S. Patent No. 11,402,015 ("the '015 Patent"), entitled "Polarization-Diverse Grating Coupler for Silicon Photonic Transceivers."
14.  With respect to the '562 Patent, I wish to provide the following additional background. I am a co-inventor on the '562 Patent, along with Dr. Patricia Emerson and Mr. Karim Nassar, both of whom were employees of Solara Optics, Inc. ("Solara Optics") at the time of invention. The '562 Patent was filed on November 22, 2014, and issued on June 13, 2017. The '562 Patent was originally assigned to Solara Optics, Inc. At the time the invention was conceived and reduced to practice, I was serving as a technical consultant to Solara Optics pursuant to a consulting agreement between OWS and Solara Optics, under which OWS provided certain engineering consulting services to Solara Optics in the area of thermal management for photonic integrated circuits. As a consultant, I made substantial contributions to the conception of the invention disclosed in the '562 Patent and was properly named as a co-inventor.
15.  On June 15, 2018, OWS acquired the '562 Patent, along with eleven other United States patents and related foreign counterparts, from Solara Optics pursuant to an Asset Purchase Agreement dated June 15, 2018, for a total purchase price of $34.5 million. Solara Optics was subsequently dissolved in 2020. As a result of the Asset Purchase Agreement, OWS is the current owner by assignment of the '562 Patent and holds all right, title, and interest therein.
C. Publications and Industry Recognition
16.  I am the author or co-author of over forty peer-reviewed publications in leading journals and conference proceedings in the field of silicon photonics and optical communications, including publications in IEEE Journal of Lightwave Technology, IEEE Photonics Technology Letters, Optics Express (Optica Publishing Group, formerly the Optical Society of America), and conference proceedings of the Optical Fiber Communication Conference (OFC), the European Conference on Optical Communications (ECOC), and the IEEE International Conference on Group IV Photonics.
17.  I have served on the program committees for the Optical Fiber Communication Conference (OFC) in 2018, 2020, and 2022, and for the IEEE Photonics Conference in 2019 and 2021. I have been invited to present keynote and plenary talks at multiple industry conferences on the topics of silicon-photonics transceiver technology and photonic integrated circuits for hyperscale data-center interconnects.
18.  I am widely recognized within the silicon-photonics industry as a leader in coherent transceiver technology and photonic integrated circuit architecture for high-bandwidth data-center interconnect applications. OWS's technology platform, which I have helped design and develop over the past fifteen years, is widely regarded as one of the leading silicon-photonics platforms for hyperscale cloud-computing applications.
D. Role in Prosecution of the Asserted Patents
19.  Throughout the prosecution of each of the four asserted patents before the United States Patent and Trademark Office ("USPTO"), I worked directly with patent prosecution counsel at Cascade IP Advisors, LLC. The lead prosecution attorney responsible for all four patents was Dr. Martin Chu, a registered patent attorney and former photonics engineer. I collaborated closely with Dr. Chu and his team during every significant phase of the prosecution process.
20.  For the '847 Patent, the '193 Patent, and the '015 Patent, my involvement in the prosecution process included: (a) preparing and providing detailed technical disclosures describing the inventions to Dr. Chu and his team; (b) reviewing and providing input on draft patent specifications and claims prior to filing; (c) reviewing and providing technical guidance on responses to office actions issued by the USPTO; (d) participating in examiner interviews with Dr. Chu, during which I provided technical explanations of the inventions and their distinctions over cited prior art; and (e) reviewing and approving claim amendments proposed by Dr. Chu in response to office actions.
21.  For the '562 Patent, my involvement was somewhat different due to my role as a consultant to Solara Optics at the time. Specifically, I provided the original technical disclosure describing the micro-channel heat-sink invention to Solara Optics' patent counsel, who was also Dr. Martin Chu of Cascade IP Advisors, LLC. I participated in the initial drafting of the patent specification and claims and provided input during the prosecution of the '562 Patent before its assignment to OWS. After OWS acquired the '562 Patent in June 2018, I reviewed the complete prosecution history as part of the due-diligence process associated with the Asset Purchase Agreement and am fully familiar with the entirety of the prosecution record.
III. OVERVIEW OF THE ASSERTED PATENTS AND THEIR TECHNOLOGICAL CONTEXT
A. The Field of Silicon Photonics for Data-Center Interconnects
22.  Before addressing the specific patents and claim terms at issue, I believe it is helpful to provide a brief overview of the field of silicon photonics and the technological context in which the asserted patents were conceived and developed.
23.  Silicon photonics is a technology that uses silicon — the same semiconductor material that forms the foundation of modern electronic integrated circuits — as an optical medium for guiding, modulating, detecting, and multiplexing light signals on a semiconductor chip. By leveraging the mature CMOS (complementary metal-oxide-semiconductor) fabrication infrastructure that has been developed over decades for electronic integrated circuits, silicon photonics enables the cost-effective mass production of optical devices on silicon wafers in high-volume semiconductor foundries.
24.  The primary commercial application of silicon photonics today is in optical transceiver modules for data-center interconnects. As the demand for bandwidth in hyperscale cloud-computing environments has grown exponentially — driven by applications such as cloud computing, artificial intelligence, machine learning, and streaming media — the need for high-bandwidth, low-power, compact optical transceiver modules has become critical. Silicon-photonic transceiver modules address this need by integrating multiple photonic functions — including laser modulation, wavelength multiplexing, photodetection, and signal processing — onto a single photonic integrated circuit ("PIC") fabricated on a silicon substrate.
25.  At the time of the earliest filing date among the asserted patents — November 22, 2014, the filing date of the '562 Patent — the silicon-photonics industry was undergoing a significant transition. The industry was moving away from traditional optical transceiver designs that used discrete optical components (such as separate lasers, modulators, and detectors assembled in a package) toward photonic integrated circuits that combine multiple optical functions on a single silicon substrate. This transition promised to reduce cost, improve performance, decrease power consumption, and enable smaller form factors, but it also introduced new technical challenges.
26.  The four asserted patents in this action each address a critical technical challenge that arose during this transition to highly integrated silicon-photonic transceivers:
(a) The '847 Patent addresses the challenge of low insertion loss in waveguide multiplexing. As the number of wavelength channels increases in a wavelength-division multiplexing (WDM) transceiver, the optical loss introduced by the multiplexer becomes a critical performance bottleneck. The '847 Patent discloses a novel waveguide multiplexer architecture that achieves significantly lower insertion loss than prior arrayed waveguide grating (AWG) designs, enabling higher-channel-count WDM transceivers.
(b) The '193 Patent addresses the challenge of integration of coherent transceiver functions on a single silicon substrate. Coherent optical transmission, which uses the phase and amplitude of light to encode data, enables significantly higher data rates than traditional intensity-modulation/direct-detection (IM/DD) approaches, but requires complex transceiver architectures that were historically implemented using discrete or hybrid-integrated components. The '193 Patent discloses a coherent optical transceiver architecture that integrates modulator, detector, and signal-processing functions on a single silicon substrate, achieving high bandwidth suitable for 400G and beyond data rates.
(c) The '562 Patent addresses the challenge of thermal management of densely integrated PICs. As photonic integrated circuits become more densely integrated, localized heat generation from active components (such as modulators, lasers, and heaters for tuning) can degrade device performance and reliability. The '562 Patent discloses a method of embedding micro-channel heat sinks within the PIC substrate to efficiently dissipate heat without increasing the device footprint.
(d) The '015 Patent addresses the challenge of efficient coupling of light into silicon waveguides with polarization diversity. Optical fibers carry light in two orthogonal polarization states, but silicon waveguides are inherently polarization-sensitive. For coherent transceiver applications, it is essential to efficiently couple and separately process both polarization states. The '015 Patent discloses a polarization-diverse grating coupler that efficiently couples light of both orthogonal polarization states into silicon waveguides.
27.  These four inventions, taken together, represent key innovations that enabled the development of high-performance, fully integrated silicon-photonic coherent transceiver modules — precisely the type of advanced transceiver modules that are now critical for hyperscale data-center interconnects operating at 400G, 800G, and beyond.
B. The '847 Patent — U.S. Patent No. 10,234,847
28.  The '847 Patent is entitled "Multi-Channel Silicon Photonic Waveguide Multiplexer with Low Insertion Loss." It was filed on September 14, 2016, and issued on March 19, 2019. The original assignee is Optiwave Semiconductor Corp. I am a named inventor on this patent, along with Dr. Sarah Chen, Co-Founder and Vice President of Engineering at OWS.
29.  The '847 Patent discloses and claims a novel waveguide multiplexer design that combines multiple optical signals of different wavelengths onto a single output waveguide with significantly reduced insertion loss compared to prior arrayed waveguide grating (AWG) designs. The key innovation lies in the multiplexer's architecture, which achieves broadband, low-loss wavelength combining through a specific arrangement of waveguide structures that is fundamentally different from conventional AWG approaches. The specification describes multiple embodiments of the waveguide multiplexer, including AWG-type structures, echelle grating structures, and cascaded Mach-Zehnder interferometer structures, all of which fall within the scope of the claimed "waveguide multiplexer."
30.  The '847 Patent is in the same patent family as the '562 Patent. Specifically, the '847 Patent is a continuation-in-part of the application that issued as the '562 Patent, sharing a common specification with the '562 Patent for certain foundational disclosures while adding new matter relating to the waveguide multiplexer innovations.
C. The '193 Patent — U.S. Patent No. 10,891,193
31.  The '193 Patent is entitled "High-Bandwidth Coherent Optical Transceiver Integrated on Silicon Substrate." It was filed on April 7, 2017, and issued on January 12, 2021. The original assignee is Optiwave Semiconductor Corp. I am a named inventor on this patent, along with Dr. Michael Faraday Okonkwo and Dr. Sarah Chen.
32.  The '193 Patent claims priority to U.S. Provisional Application No. 62/499,117, filed April 7, 2017. This provisional application is shared with the '015 Patent, as described below.
33.  The '193 Patent discloses and claims a coherent optical transceiver architecture that integrates modulator, detector, and signal-processing functions on a single silicon substrate. The transceiver achieves high bandwidth suitable for 400G and beyond data rates through novel circuit topologies that can be fabricated using standard CMOS-compatible processes. The claimed architecture includes a coherent detection module that performs coherent detection of incoming optical signals, including mixing the incoming signal with a local oscillator and detecting the resulting interference pattern. As used in the claims, the term "coherent detection module" refers broadly to the functional module that performs coherent detection, and is not limited to any single specific hardware implementation.
D. The '562 Patent — U.S. Patent No. 9,678,562
34.  The '562 Patent is entitled "Method for Thermal Management of Photonic Integrated Circuits Using Embedded Micro-Channel Heat Sinks." It was filed on November 22, 2014, and issued on June 13, 2017. As discussed above, the '562 Patent was originally assigned to Solara Optics, Inc. and is currently assigned to OWS pursuant to the Asset Purchase Agreement dated June 15, 2018.
35.  I am a co-inventor on the '562 Patent, along with Dr. Patricia Emerson and Mr. Karim Nassar. The '562 Patent is in the same patent family as the '847 Patent, as described above.
36.  The '562 Patent discloses and claims a method of embedding micro-channel heat sinks within photonic integrated circuits to manage thermal dissipation in densely integrated devices. The embedded micro-channel structures allow for efficient, localized heat removal from active photonic components without increasing the overall device footprint. The specification describes several embodiments of the micro-channel heat-sink architecture, including embodiments with different channel geometries, different coolant delivery configurations (including both active and passive cooling approaches), and different integration methods for embedding the micro-channels within the PIC substrate.
E. The '015 Patent — U.S. Patent No. 11,402,015
37.  The '015 Patent is entitled "Polarization-Diverse Grating Coupler for Silicon Photonic Transceivers." It was filed on August 30, 2019, and issued on August 2, 2022. The original assignee is Optiwave Semiconductor Corp. I am a named inventor on this patent, along with Dr. Ananya Krishnamurthy and Dr. Michael Faraday Okonkwo.
38.  The '015 Patent claims priority to U.S. Provisional Application No. 62/499,117, filed April 7, 2017 — the same provisional application to which the '193 Patent claims priority. The shared priority date and common provisional disclosure demonstrate that the coherent transceiver architecture of the '193 Patent and the polarization-diverse grating coupler of the '015 Patent were conceived as complementary innovations within a unified silicon-photonic transceiver architecture.
39.  The '015 Patent discloses and claims a grating coupler design that efficiently couples light of two orthogonal polarization states (commonly referred to as TE and TM polarization modes) into silicon waveguides, enabling polarization diversity in transceiver architectures. The specification describes multiple embodiments for achieving polarization diversity, including two-dimensional grating structures, one-dimensional grating couplers paired with polarization rotator-splitters, and grating couplers with integrated sub-wavelength polarization-diverse elements. The independent claims use the broad term "polarization-diverse grating coupler" to encompass all of these embodiments, while dependent claims — such as Claim 12, which recites a "two-dimensional sub-wavelength grating" — add specific structural limitations corresponding to particular embodiments.
IV. STATE OF THE ART AT THE TIME OF EACH PATENT FILING
A. State of the Art at the Time of the '562 Patent Filing (November 2014)
40.  At the time the '562 Patent was filed on November 22, 2014, thermal management in photonic integrated circuits was widely recognized as a critical challenge facing the silicon-photonics industry. As PICs became more densely integrated — incorporating more active components such as modulators, tunable filters, and heater elements on a single chip — the localized heat generated by these components created thermal "hot spots" that degraded device performance, shifted operating wavelengths, and reduced long-term reliability.
41.  The existing approaches to thermal management in PICs at the time of filing relied primarily on bulk heat-spreading solutions or thermoelectric coolers (TECs) that were external to the photonic circuit itself. Bulk heat spreaders — typically copper or aluminum blocks attached to the bottom of the PIC package — provided only coarse-grained thermal management and were unable to address localized hot spots within the PIC. Thermoelectric coolers could maintain the entire PIC at a uniform temperature, but consumed significant power and added to the package size and cost.
42.  The concept of embedding micro-channel structures directly within the PIC substrate for localized thermal management was not well developed at the time of the '562 Patent's filing. Prior approaches to micro-channel cooling existed in the broader field of electronic integrated circuits (dating back to the seminal work of Tuckerman and Pease at Stanford in 1981), but these approaches had not been successfully adapted to the unique constraints of photonic integrated circuits, which require precise optical alignment and are sensitive to mechanical stresses that can alter waveguide properties. Some researchers had explored macro-scale liquid-cooling solutions for optical systems, but these approaches used external cooling loops and were not integrated into the PIC substrate. Passive heat-sink solutions that were available at the time could not adequately address the localized hot spots generated by densely integrated active photonic components.
43.  The '562 Patent addressed a long-felt need in the silicon-photonics industry for an effective thermal management solution that could be integrated directly into the PIC substrate without increasing the device footprint or compromising optical performance. The embedded micro-channel heat-sink architecture disclosed in the '562 Patent provided a means for targeted, localized heat removal from the most thermally critical regions of the PIC.
44.  A person of ordinary skill in the art ("POSITA") at the time of the '562 Patent's filing would have been a person with at least a Master of Science degree in electrical engineering, photonics, or a closely related field, and at least three years of professional experience in the design, fabrication, or testing of silicon-photonic integrated circuits. This definition is consistent with the POSITA definition proposed by OWS's technical expert, Dr. Margaret Ashford, in her separate expert declaration filed in support of OWS's Opening Claim Construction Brief.
B. State of the Art at the Time of the '847 Patent Filing (September 2016)
45.  At the time the '847 Patent was filed on September 14, 2016, arrayed waveguide gratings (AWGs) were the predominant technology for wavelength-division multiplexing in silicon photonics. AWGs use an array of waveguides with precisely controlled path-length differences to create constructive interference at specific output ports for specific wavelengths, enabling the combining (multiplexing) or separating (demultiplexing) of multiple wavelength channels. However, AWGs suffered from inherently high insertion loss, particularly as the number of wavelength channels increased. For example, a typical 8-channel AWG designed for CWDM (coarse wavelength-division multiplexing) applications in silicon photonics at that time exhibited insertion losses on the order of 3-5 dB per channel, which became a significant performance bottleneck as channel counts increased to 16 or more channels for DWDM (dense wavelength-division multiplexing) applications.
46.  The '847 Patent's innovation was a novel waveguide multiplexer architecture that achieved significantly lower insertion loss than conventional AWGs while maintaining a compact form factor suitable for photonic integrated circuit applications. The claimed waveguide multiplexer was not limited to any particular multiplexing architecture; rather, the patent disclosed a set of design principles and structural features that could be implemented using various waveguide-based multiplexing approaches, including but not limited to echelle grating structures and cascaded Mach-Zehnder interferometer structures, in addition to improved AWG designs.
47.  Other researchers in the field had explored silicon photonic wavelength multiplexers around this time. Notably, Takahashi et al. published an article entitled "Dense WDM Silicon Photonic MUX Using Higher-Order Microring Resonators" in IEEE Photonics Technology Letters in March 2014, which described a dense WDM silicon photonic multiplexer based on an AWG architecture with a slab coupler. However, Takahashi's approach did not achieve the specific combination of low insertion loss and scalability to high channel counts that is claimed in the '847 Patent. Takahashi's design was fundamentally limited by the conventional AWG architecture and its inherent loss mechanisms, including slab-coupler diffraction loss and waveguide-array non-uniformity. The '847 Patent's approach differed fundamentally from Takahashi's AWG-based architecture in its use of multiplexing structures that are not limited to arrayed waveguide gratings and that achieve low loss through different structural mechanisms.
C. State of the Art at the Time of the '193 Patent Filing (April 2017)
48.  At the time the '193 Patent was filed on April 7, 2017, coherent optical transceivers were emerging as an essential technology for achieving data rates of 400G and beyond in data-center interconnect applications. Coherent modulation formats, such as DP-QPSK (dual-polarization quadrature phase-shift keying) and DP-16QAM (dual-polarization 16-state quadrature amplitude modulation), offered significantly higher spectral efficiency than the intensity-modulation/direct-detection (IM/DD) approaches that had been used for earlier-generation 100G transceivers. However, the integration of all coherent transceiver functions — including the transmitter modulator, the local oscillator, the coherent receiver (including 90-degree optical hybrids and balanced photodetectors), and associated electronic signal processing — on a single silicon substrate had not yet been achieved.
49.  Prior coherent transceiver designs relied on hybrid integration approaches, in which III-V semiconductor dies (such as indium phosphide-based lasers, modulators, and photodetectors) were physically bonded to a silicon photonic die using techniques such as flip-chip bonding or micro-transfer printing. While hybrid integration leveraged the superior optical gain properties of III-V semiconductors, it introduced significant challenges, including precise alignment requirements between the III-V and silicon dies, increased manufacturing cost, reduced yield, and potential reliability concerns due to the heterogeneous materials interfaces.
50.  The '193 Patent achieved what I describe as "monolithic-like" integration — a level of integration that approached the simplicity and reliability of monolithic fabrication — through novel circuit topologies that could be implemented using standard CMOS-compatible fabrication processes. The '193 Patent's architecture enabled the integration of coherent transceiver functions on a single silicon substrate while achieving high bandwidth, low power consumption, and compact footprint. Importantly, the claims of the '193 Patent recite integration "on a silicon substrate," which encompasses both monolithic fabrication and heterogeneous integration approaches that use a silicon wafer or die as the primary platform.
51.  A POSITA at the time of the '193 Patent's filing would have understood the phrase "integrated on a silicon substrate" to refer to any fabrication approach in which the functional components of the transceiver are formed on or bonded to a silicon-based wafer, using semiconductor fabrication processes. This understanding was consistent with the way the phrase was used throughout the silicon-photonics industry at that time, which recognized both monolithic silicon photonics and heterogeneous integration on silicon as forms of "integration on a silicon substrate."
D. State of the Art at the Time of the '015 Patent Filing (August 2019)
52.  At the time the '015 Patent was filed on August 30, 2019, grating couplers were a standard and widely used method for coupling light between optical fibers and silicon waveguides in photonic integrated circuits. Grating couplers offer the advantage of enabling vertical or near-vertical coupling of light from a fiber into a planar waveguide, which simplifies packaging and testing compared to edge-coupling approaches. However, most existing grating coupler designs at the time of filing were polarization-sensitive — that is, they were designed to efficiently couple only one polarization state (typically the TE mode) while rejecting or poorly coupling the orthogonal polarization state (the TM mode). This polarization sensitivity was a significant limitation for coherent transceiver applications, which require the processing of both orthogonal polarization states to implement dual-polarization modulation formats.
53.  Two-dimensional (2D) grating couplers had been proposed in the academic literature as a potential solution for achieving polarization diversity. A 2D grating coupler uses a grating pattern that is periodic in two dimensions, allowing it to couple TE-polarized light into one output waveguide and TM-polarized light into a different output waveguide. However, prior 2D grating coupler designs suffered from low coupling efficiency (typically less than -5 dB) and narrow optical bandwidth, limiting their practical utility in high-performance transceiver applications.
54.  The '015 Patent's innovation was a polarization-diverse grating coupler that could efficiently couple both TE and TM polarization modes with high efficiency across a broad wavelength range suitable for WDM transceiver applications. The patent disclosed multiple structural approaches for achieving polarization diversity, as described in paragraph 39 above. The "two-dimensional sub-wavelength grating" limitation recited in dependent Claim 12 refers to a specific structural feature — the use of sub-wavelength patterning within a two-dimensional grating to enhance coupling efficiency and broaden bandwidth — that corresponds to one of the disclosed embodiments. However, the broader concept of "polarization-diverse grating coupler" used in the independent claims encompasses any grating coupler that is capable of coupling light of two orthogonal polarization states, regardless of the specific structural mechanism employed.
V. PROSECUTION HISTORY AND INTENDED SCOPE OF DISPUTED CLAIM TERMS
55.  I turn now to the four disputed claim terms identified in the parties' Joint Claim Construction and Prehearing Statement. For each term, I provide my testimony regarding the intended scope of the term as understood by me as an inventor, the relevant prosecution history, and the understanding that a POSITA would have had of the term at the time of filing.
A. "Waveguide Multiplexer" ('847 Patent, Claim 1)
56.  OWS's proposed construction: "An optical device that combines multiple optical signals of different wavelengths onto a single waveguide."
57.  Lumicore's proposed construction: "An arrayed waveguide grating that combines optical signals using a slab coupler and phase-shifted waveguide array."
58.  The term "waveguide multiplexer" was deliberately chosen during the drafting of the '847 Patent to encompass a broad range of optical multiplexing devices that use waveguide structures to combine multiple optical signals of different wavelengths. The term was not intended to be limited to any particular multiplexing architecture, including but not limited to arrayed waveguide gratings (AWGs). I was directly involved in the selection of this claim language, and I can attest that the decision to use the broad term "waveguide multiplexer" rather than a more specific architectural term (such as "arrayed waveguide grating," "echelle grating multiplexer," or "cascaded Mach-Zehnder interferometer multiplexer") was a deliberate and informed choice.
59.  The specification of the '847 Patent fully supports this broad construction. The specification describes multiple embodiments of the waveguide multiplexer, including:
(a) An AWG-type multiplexer using a slab coupler and an array of waveguides with controlled path-length differences (described at col. 5, ll. 12-45 of the '847 Patent);
(b) An echelle grating multiplexer using a reflective concave grating etched into the silicon substrate (described at col. 7, ll. 3-38 of the '847 Patent); and
(c) A cascaded Mach-Zehnder interferometer (MZI) multiplexer using a binary-tree arrangement of MZI stages (described at col. 8, ll. 50 through col. 9, l. 22 of the '847 Patent).
60.  Moreover, the specification of the '847 Patent explicitly states, at column 4, lines 56-58: "The waveguide multiplexer of the present invention is not limited to any particular optical combining architecture." This statement was included in the specification at my specific direction during the drafting process, precisely to ensure that the claims would not be improperly limited to any single embodiment.
61.  Lumicore's proposed construction — which limits "waveguide multiplexer" to "an arrayed waveguide grating that combines optical signals using a slab coupler and phase-shifted waveguide array" — improperly imports limitations from one specific embodiment described in the specification into the independent claims. This proposed construction would exclude two of the three embodiments expressly described in the specification (the echelle grating multiplexer and the cascaded MZI multiplexer), which is contrary to well-established claim construction principles.
62.  The prosecution history of the '847 Patent further supports OWS's proposed construction. During prosecution, the examiner initially rejected certain claims of the '847 Patent over prior art AWG references, including the Takahashi et al. reference discussed above. In response, Dr. Chu and I distinguished the claimed waveguide multiplexer from the specific prior art AWG designs by pointing to the structural features described in the dependent claims — not by limiting the independent claim term "waveguide multiplexer" to any particular architecture. Specifically, in the Office Action Response filed on June 18, 2018, we distinguished Takahashi by arguing that Takahashi's AWG design did not achieve the specific insertion-loss performance recited in dependent Claim 3 and did not incorporate the specific tapered waveguide transition structure recited in dependent Claim 5. At no point in this response, or in any other filing during prosecution, did we limit the term "waveguide multiplexer" in the independent claims to an AWG architecture.
63.  The examiner allowed the claims after considering our response, without requiring any narrowing amendment to the term "waveguide multiplexer" in the independent claims. The examiner's reasons for allowance, as stated in the Notice of Allowability, referenced the combination of features recited in independent Claim 1 as a whole, not any particular architectural limitation on the term "waveguide multiplexer."
64.  A POSITA at the time of filing would have understood the term "waveguide multiplexer" to refer broadly to any optical device that combines multiple optical signals of different wavelengths onto a single waveguide using waveguide-based structures. The term "waveguide multiplexer" is a commonly used term in the silicon-photonics literature and is not understood in the art to be synonymous with "arrayed waveguide grating." AWGs are one type of waveguide multiplexer, but the term encompasses other architectures as well.
B. "Integrated on a Silicon Substrate" ('193 Patent, Claim 1)
65.  OWS's proposed construction: "Formed on or within a silicon-based wafer using semiconductor fabrication processes."
66.  Lumicore's proposed construction: "Monolithically fabricated on a single silicon die without hybrid bonding or heterogeneous integration."
67.  The phrase "integrated on a silicon substrate" was used in the claims of the '193 Patent to distinguish the claimed coherent optical transceiver from free-space optical systems or fiber-based optical systems that do not involve integration of components on a semiconductor substrate. The phrase was not intended to exclude heterogeneous integration techniques that use a silicon substrate as the base platform. I was directly involved in the drafting of this claim language, and the decision to use the phrase "integrated on a silicon substrate" — rather than a more restrictive phrase such as "monolithically fabricated on a single silicon die" — was deliberate.
68.  The specification of the '193 Patent confirms that "integrated on a silicon substrate" is not limited to monolithic fabrication. The specification describes multiple embodiments of the coherent transceiver architecture, including embodiments where certain components — specifically, the laser source and, in some embodiments, certain high-speed photodetectors — are fabricated using III-V semiconductor materials (such as indium phosphide or gallium arsenide) and then bonded to the silicon substrate as part of the fabrication process. The specification explicitly describes these hybrid configurations as being "integrated on a silicon substrate." For example, at column 6, lines 22-28, the specification states: "In certain embodiments, the laser source 110 comprises an indium phosphide gain chip that is heterogeneously integrated on the silicon substrate 100 using wafer-level bonding techniques. Such heterogeneous integration is within the scope of the present invention, as the silicon substrate serves as the primary platform on which all transceiver functions are integrated."
69.  During prosecution of the '193 Patent before the USPTO, the examiner did not require the applicant to disclaim heterogeneous integration. No amendment was made during prosecution that narrowed the phrase "integrated on a silicon substrate" in any way. The prosecution history is entirely silent on any distinction between monolithic and heterogeneous integration, which further confirms that the phrase was understood broadly by both the applicant and the examiner.
70.  The state of the art in April 2017, at the time of the '193 Patent's filing, recognized that "integration on a silicon substrate" encompassed both monolithic and heterogeneous integration approaches, as long as the silicon wafer or die served as the primary platform for the integrated device. The term "silicon photonics" itself, as used in the industry at that time, encompassed devices that used heterogeneous integration of III-V components on silicon substrates. Major research groups and companies in the field — including Intel, Greaves Holdings (formerly Crestline Corporation), and Brightmoor Industries — routinely described their heterogeneously integrated silicon-photonic devices as being "integrated on a silicon substrate" or "integrated on a silicon photonic platform."
71.  A POSITA at the time of filing would have understood the phrase "integrated on a silicon substrate" to include any fabrication approach where the functional components of the transceiver are formed on or bonded to a silicon-based wafer using semiconductor fabrication processes. Lumicore's proposed construction, which excludes heterogeneous integration, is contrary to the specification, the prosecution history, and the understanding of a POSITA at the time of filing.
C. "Micro-Channel Heat Sink" ('562 Patent, Claim 1)
72.  OWS's proposed construction: "A heat-dissipating structure containing channels with a hydraulic diameter of less than 1 mm."
73.  Lumicore's proposed construction: "A closed-loop liquid-cooling system with channels having a hydraulic diameter between 10 µm and 200 µm."
74.  The specification of the '562 Patent uses the term "micro-channel" to refer to channels with dimensions in the micrometer range, but does not impose an upper limit of 200 µm as Lumicore contends. Rather, the specification provides examples of channels with hydraulic diameters ranging from approximately 50 µm to approximately 800 µm. Specifically, the specification states at column 4, lines 33-38: "The micro-channels 210 have hydraulic diameters ranging from approximately 50 µm to approximately 800 µm, depending on the thermal dissipation requirements and the available substrate thickness of the photonic integrated circuit." The specification further states at column 5, lines 10-14: "In a preferred embodiment, the micro-channels have a hydraulic diameter of approximately 100 µm, but the invention is not limited to any particular channel dimension within the micrometer range."
75.  Lumicore's attempt to impose an upper limit of 200 µm on the hydraulic diameter of the micro-channels has no support in the specification or the prosecution history. Lumicore appears to be relying on a single sentence in the specification that describes one preferred embodiment as having channels "with hydraulic diameters between approximately 50 µm and 200 µm" (col. 5, ll. 20-22). However, this sentence describes a single preferred embodiment and cannot be used to limit the scope of the claims to that particular embodiment, especially when the specification explicitly describes embodiments with channels having hydraulic diameters up to approximately 800 µm.
76.  Lumicore's proposed construction also improperly requires a "closed-loop liquid-cooling system." The specification of the '562 Patent does not require a closed-loop configuration. To the contrary, the specification describes several embodiments with different coolant delivery configurations:
(a) A closed-loop liquid-cooling configuration, in which a coolant (such as deionized water or a dielectric fluid) is circulated through the micro-channels by a micropump and returned to a reservoir (described at col. 6, ll. 5-25);
(b) A single-pass or open-loop configuration, in which coolant flows through the micro-channels in a single pass without recirculation (described at col. 6, ll. 26-40); and
(c) A passive two-phase cooling configuration, in which a working fluid undergoes phase change (evaporation and condensation) within the micro-channels to transfer heat, without an active pumping system (described at col. 7, ll. 1-22).
77.  All three of these configurations are described in the specification as embodiments of the "micro-channel heat sink." Lumicore's proposed construction would exclude the single-pass and passive two-phase cooling embodiments, which are expressly described in the specification. This is improper.
78.  The prosecution history of the '562 Patent supports OWS's proposed construction. During prosecution, the examiner issued an office action citing the conference paper by Dr. Li Wei, entitled "Thermal Solutions for PICs," presented at the Optical Fiber Communication Conference (OFC) in 2013. In our response, Dr. Chu and I distinguished Dr. Li Wei's approach by pointing to specific structural differences in the micro-channel geometry and integration method disclosed in the '562 Patent — specifically, the fact that the '562 Patent's micro-channels are embedded within the PIC substrate itself, whereas Dr. Li Wei's approach used external cooling structures attached to the outside of the PIC package. At no point during this response, or at any other time during prosecution, did we limit the term "micro-channel heat sink" to a specific hydraulic diameter range or to a closed-loop cooling-loop architecture.
79.  The term "micro-channel" is a well-established term of art in the thermal management field. The seminal paper by Tuckerman and Pease, published in 1981, defined micro-channels as channels with hydraulic diameters on the order of micrometers, and subsequent literature has generally used the term to refer to channels with hydraulic diameters below approximately 1 mm. This is consistent with OWS's proposed construction and with the range of channel dimensions described in the specification of the '562 Patent.
80.  A POSITA at the time of filing would have understood the term "micro-channel heat sink" as used in the '562 Patent to mean a heat-dissipating structure containing channels with a hydraulic diameter of less than 1 mm. A POSITA would not have understood this term to require a closed-loop liquid-cooling system or to impose an upper limit of 200 µm on the channel hydraulic diameter, as Lumicore proposes. Lumicore's proposed construction improperly imports limitations from specific preferred embodiments and from its own preferred design choices into the claims.
D. "Polarization-Diverse Grating Coupler" ('015 Patent, Claim 1)
81.  OWS's proposed construction: "A grating coupler capable of coupling light of two orthogonal polarization states."
82.  Lumicore's proposed construction: "A two-dimensional grating coupler with a polarization-splitting element that simultaneously couples TE and TM modes."
83.  The independent claims of the '015 Patent use the term "polarization-diverse grating coupler" without requiring that the coupler be two-dimensional or that it include a polarization-splitting element as a separate structural component. The term was selected during drafting to encompass the full range of grating coupler designs that achieve polarization diversity — that is, the ability to couple light of two orthogonal polarization states (TE and TM) into silicon waveguides. I was directly involved in the selection of this claim language.
84.  The specification of the '015 Patent describes multiple embodiments of the polarization-diverse grating coupler, including:
(a) A two-dimensional grating coupler that inherently splits polarizations by directing TE-polarized light into one output waveguide and TM-polarized light into a different output waveguide, based on the two-dimensional periodicity of the grating structure (described at col. 4, ll. 15-45);
(b) A one-dimensional grating coupler paired with a separate polarization rotator-splitter, where the grating coupler first couples both polarization states into a single waveguide, and a downstream polarization rotator-splitter then separates the two polarization states into separate waveguides (described at col. 6, ll. 8-35); and
(c) A one-dimensional grating coupler with an integrated polarization-diverse structure using sub-wavelength elements, where the sub-wavelength patterning within the grating itself provides polarization-selective coupling without requiring a separate polarization-splitting element (described at col. 7, ll. 10-42).
85.  All three of these embodiments are described in the specification as implementations of a "polarization-diverse grating coupler." Lumicore's proposed construction — which requires a "two-dimensional grating coupler with a polarization-splitting element" — would improperly exclude the one-dimensional grating coupler embodiments described in embodiments (b) and (c) above, which are expressly disclosed in the specification.
86.  The principle of claim differentiation provides further support for OWS's broader construction. Dependent Claim 12 of the '015 Patent specifically recites "wherein the polarization-diverse grating coupler comprises a two-dimensional sub-wavelength grating." The inclusion of this specific structural limitation in a dependent claim demonstrates that the drafters — Dr. Chu and I — intended the independent claims, which do not include the "two-dimensional sub-wavelength grating" limitation, to be broader than Claim 12. If Lumicore's proposed construction were adopted — requiring a two-dimensional grating coupler in the independent claims — it would render the additional limitation in dependent Claim 12 superfluous, contrary to the principle of claim differentiation.
87.  The prosecution history confirms this analysis. During prosecution, the examiner issued a rejection of Claim 12 under 35 U.S.C. § 103, based on a combination of the Bergmann patent (U.S. Patent No. 8,412,007) and the Simonsen PCT application (PCT/EP2015/071234). In response to this rejection, Dr. Chu and I amended Claim 12 — and only Claim 12 — to add the "sub-wavelength" qualifier to the "two-dimensional grating" limitation. The amendment changed Claim 12 from reciting "a two-dimensional grating" to reciting "a two-dimensional sub-wavelength grating." This amendment was specifically targeted at the Bergmann-Simonsen combination and was limited to dependent Claim 12. The independent claims were not amended, and the term "polarization-diverse grating coupler" in the independent claims was not modified in any way. The examiner subsequently allowed all claims, including the independent claims, without requiring any amendment to the term "polarization-diverse grating coupler."
88.  A POSITA at the time of filing would have understood the term "polarization-diverse grating coupler" to mean any grating coupler designed to handle both orthogonal polarization states of incoming light, regardless of the specific structural mechanism used to achieve polarization diversity. The term "polarization-diverse" is a well-established modifier in the optical communications field that indicates the capability of processing or handling both polarization states, without implying any particular structural implementation.
VI. ADDITIONAL OBSERVATIONS ON CLAIM TERMS AND PROSECUTION HISTORY
A. Continuation Relationship Between the '562 and '847 Patents
89.  As discussed above, the '562 Patent and the '847 Patent are in the same patent family — the '847 Patent is a continuation-in-part of the application that issued as the '562 Patent. Because these two patents share common specification language for certain foundational disclosures, claim terms that are common to both patents should be construed consistently across the two patents. The prosecution histories of both patents were coordinated by the same patent attorney, Dr. Martin Chu of Cascade IP Advisors, LLC, and I was personally involved in the prosecution of both patents. The same technical terminology was used with the same intended meaning in both patent specifications and claims.
B. Common Provisional Application for the '193 and '015 Patents
90.  Both the '193 Patent and the '015 Patent claim priority to Provisional Application No. 62/499,117, filed on April 7, 2017. The provisional application, which I co-authored, contains disclosure supporting both the coherent transceiver architecture claimed in the '193 Patent and the polarization-diverse grating coupler claimed in the '015 Patent. The provisional application describes the polarization-diverse grating coupler as a key component of the coherent transceiver architecture, functioning as the input coupler that receives incoming optical signals from the fiber and delivers them, separated by polarization, to the coherent receiver circuitry integrated on the silicon substrate.
91.  The shared priority date and common provisional disclosure demonstrate that the coherent transceiver innovation of the '193 Patent and the polarization-diverse grating coupler innovation of the '015 Patent were conceived as complementary innovations within a unified silicon-photonic transceiver architecture. This common origin further supports OWS's proposed constructions, which are consistent with the way these claim terms are used throughout the provisional application and the subsequent non-provisional specifications.
C. No Disclaimer of Claim Scope During Prosecution
92.  I hereby attest that at no point during the prosecution of any of the four asserted patents did I or Dr. Chu make any statement to the USPTO that was intended to, or that should be construed to, disclaim the scope of any of the disputed independent claim terms as proposed by OWS in its Opening Claim Construction Brief. The amendments that were made during prosecution were narrowly targeted at specific dependent claims and specific prior-art references, and were not intended to narrow the scope of the independent claim terms that are the subject of the current claim construction dispute.
93.  Specifically:
(a) For the '847 Patent, the distinction over Takahashi et al. was made by pointing to structural features in the dependent claims, not by narrowing "waveguide multiplexer" in the independent claims.
(b) For the '193 Patent, no amendment was made during prosecution that addressed or narrowed the phrase "integrated on a silicon substrate."
(c) For the '562 Patent, the distinction over the Li Wei conference paper was based on the specific structural differences between the claimed embedded micro-channel architecture and Li Wei's external thermal management approach, not on limiting the definition of "micro-channel heat sink."
(d) For the '015 Patent, the only substantive amendment during prosecution was to dependent Claim 12, adding the "sub-wavelength" qualifier. The independent claims and the term "polarization-diverse grating coupler" were not amended.
D. Consistency with OWS's Own Products
94.  OWS is one of the leading U.S. suppliers of silicon-photonic transceiver modules for hyperscale data-center interconnect applications. OWS's own product lines — the OWS-400ZR, OWS-800XR, and OWS-1600DR transceiver module families — practice the technologies claimed in the asserted patents. The way OWS uses the disputed claim terms in its internal engineering documentation, product specifications, data sheets, and marketing materials is fully consistent with the claim constructions proposed by OWS in its Opening Claim Construction Brief.
95.  For example, OWS's internal engineering documentation describes the OWS-400ZR transceiver module as incorporating a "waveguide multiplexer" for combining four wavelength channels onto a single output waveguide. Notably, the specific waveguide multiplexer used in the OWS-400ZR is an echelle-grating-based multiplexer, not an AWG. If Lumicore's unduly narrow construction of "waveguide multiplexer" — which limits the term to an AWG architecture — were adopted, OWS's own commercial embodiment of the patented technology would be excluded from the scope of the claims. This result would be inconsistent with the purpose and intent of the patent, which was drafted to cover OWS's own technology, among other implementations.
96.  Similarly, OWS's OWS-800XR product uses heterogeneous integration of an indium phosphide laser on a silicon photonic substrate, and OWS's engineering documentation describes this product as having its transceiver functions "integrated on a silicon substrate." This usage is consistent with OWS's proposed construction and inconsistent with Lumicore's proposed construction, which would exclude heterogeneous integration.
VII. DISCUSSION OF SPECIFIC PRIOR ART REFERENCES CITED IN PROSECUTION
97.  I provide the following additional detail regarding specific prior art references that were cited and considered during the prosecution of the asserted patents, as these references are relevant to the claim construction dispute.
A. Takahashi et al. ("Dense WDM Silicon Photonic MUX," IEEE Photonics Technology Letters, March 2014)
98.  The Takahashi et al. reference was cited by the examiner during prosecution of the '847 Patent. Takahashi described a dense WDM silicon photonic multiplexer based on a conventional AWG architecture employing a slab coupler and a waveguide array with precisely controlled path-length differences. While Takahashi achieved a functional wavelength multiplexer in silicon photonics, Takahashi's approach was limited to the conventional AWG architecture and suffered from the insertion-loss limitations inherent to that architecture, particularly at higher channel counts.
99.  During prosecution, Dr. Chu and I distinguished Takahashi by pointing to specific structural features of the claimed waveguide multiplexer that were recited in the dependent claims, including the specific tapered waveguide transition structures (Claim 5) and the insertion-loss performance (Claim 3). We did not, at any point, concede that the independent claim term "waveguide multiplexer" is limited to an AWG architecture. To the contrary, our arguments distinguished Takahashi on the basis of specific structural and performance features, while preserving the breadth of the independent claim terms.
100.  The examiner allowed the claims after considering our response, and the Notice of Allowability did not characterize the independent claim term "waveguide multiplexer" as being limited to any particular multiplexing architecture. Lumicore's assertion that the prosecution history supports its narrow construction of "waveguide multiplexer" is incorrect.
B. Bergmann (U.S. Patent No. 8,412,007, Issued April 2, 2013)
101.  The Bergmann patent was cited during the prosecution of both the '562 Patent and the '015 Patent.
102.  With respect to the '562 Patent, Bergmann disclosed a heat-sink structure for semiconductor devices, but did not disclose the specific concept of embedding micro-channel heat sinks directly within the substrate of a photonic integrated circuit. Bergmann's heat-sink structure was external to the semiconductor die and was not integrated into the photonic circuit itself. During prosecution, the distinction over Bergmann was made on the basis of the integration method — specifically, the embedding of micro-channels within the PIC substrate — not on the basis of narrowing the definition of "micro-channel heat sink" to a particular hydraulic diameter range or cooling-loop architecture.
103.  With respect to the '015 Paten